Fishman Fluence Classics Bode Plots

[gpdb]

I was able to measure my set of Fishman Fluence Classics (open core), and they had some really interesting results. First, Here's how I achieved it with the standard setup:



On the fishman pickup, there's an output wire that is next to the red battery wire. That wire has an output and a ground shield. I connected both parts of that wire to the integrator. I connected the red wire of the fishman to a 9V connector, and then I connected the black wire of that 9V to ground. To switch to the different voices, I just added the voice wire with the output wire on the integrator. My results are below, but note that I'm not sure how accurate these are. I'll explain more below:

Classic Bridge:


Classic Neck compared to some passive PAF-style pickups:


Classic Neck: 

 

The resonant peaks of all my charts are either exactly on or very close to what Fishman reports as the resonant peaks of each pickup. Voice 2 of the neck pickup also is described as having two resonant peaks, and is a "Fluence Exclusive" which means it's basically impossible to achieve with a real pickup. After I created these, I reached out to one of the R&D members of the Fishman team to ask if these were correct. He stated that even though the resonant peaks were in the right place, that Fluence pickups have more high end than what's being shown in these graphs. This was what he sent about what the chart should look closer to:




He also stated that the Voice 2 of the neck pickup shouldn't have nearly that much of a peak at 400hz as my graph shows. Does anyone know what I may be doing wrong to not capture these pieces correctly? Voice 1 of both look very much like a standard humbucker, so I'm not sure how one voice could look spot on and the others could be off. Maybe it's just a capability that the drive coil can't capture.

[antigua]

That is interesting. One thing to keep in mind is that the "passive PAF-style pickups" are interacting with the 470pF and 200k ohms dummy load as a hi-Z load, but the Fishman Fluence is a low-Z load, so the cable capacitance is "built in" to the response curves, so if they were identical it would be a fun coincidence, but for all intents and purposes the Fishman's classic voice is an accurate representation of a passive PAF curve.  That first plot shows some rather extreme volume differences, but aside from that, it looks like what you'd expect to see from varying guitar cable lengths and pot values with a passive pickup, ranging from no load to a standard in-situ load.

That's pretty neat that someone from Fishman answered your inquires. Ken Willmott might be able to make a claim about whether his integrator is or is not as accurate in the case of a low-Z active pickup, but my inclination is to say that if you measured X, then the reality of it is X. The concern with the driver coil is that it will inductively link with the pickup and therefore actually change the properties of the pickup, the solution to that is to have the driver coil farther away from the pickup or have the driver coil be rather small, so that the coupling coefficient will be small. If you're worried that the driver coil causes interference, just move it farther away and test again. 

The dual resonance is a nice change of pace, though for the record that can be achieved passively, for example with the Gibson Varitone control, where you have an inductor and a cap in series, parallel to the pickup, and that creates a scoop in the response curve, where the frequency of the scoop is moved around by changing cap values with a selector switch. I can tell you from experience though that a scoop around 2kHz is not he most interesting sound, because it's at a rather high frequency. If you scoop around 500Hz, you can actually make a bridge pickup sound a lot like the neck pickup of a guitar, which would be a useful more for a guitar with only a bridge pickup, but nonetheless I've never seen it done on a production guitar, possibly because the inductor required for that type of circuit is not cheap, or has to be hand made by the guitar maker. The Bill Lawrence Q filter is an example of the kind of inductor that is ideal for it. 

[aquin43]

What you are plotting through the integrator is the response to string velocity, which is the convention adopted here. If Fishman plot the amplitude response then that would explain their different expectations.

[antigua]

Jul 21, 2022 4:16:31 GMT -5 aquin43 said:

What you are plotting through the integrator is the response to string velocity, which is the convention adopted here. If Fishman plot the amplitude response then that would explain their different expectations.

If the integrator were turned off, or excluded from the test circuit, is that what you would describe as amplitude response, or are we still talking about two different things?  

[aquin43]

Jul 21, 2022 13:10:26 GMT -5 antigua said:

Jul 21, 2022 4:16:31 GMT -5 aquin43 said:

What you are plotting through the integrator is the response to string velocity, which is the convention adopted here. If Fishman plot the amplitude response then that would explain their different expectations.

If the integrator were turned off, or excluded from the test circuit, is that what you would describe as amplitude response, or are we still talking about two different things?  

The long answer:

With a string:

Flux through the coil depends on the string position.

Change in flux depends on change of string position = displacement (amplitude)
Rate of change of flux depends on rate of change of displacement = velocity

The pickup responds to rate of change of flux which maps to string velocity.

With the exciter:
Flux through the coil depends on the exciter current which maps to string position.
Change in flux depends on change of exciter current which maps to string displacement.
Rate of change of flux depends on rate of change of exciter current which maps to string velocity.

In order to measure the response of the pickup to the equivalent of a constant string velocity we need to make the rate of change of the exciter current be constant with frequency. A fixed amplitude of exciter current will give a rate of change that increases in proportion to frequency. Therefore, we need to make the exciter current fall inversely with frequency. This is difficult to do reliably so we adopt an alternative which is to use a constant exciter current and apply a correction to the reading using an integrator whose response falls inversely with frequency.

Therefore, the normal reading via the integrator represents the string velocity response and the reading without the integrator will represent the displacement or amplitude response.

[antigua]

I see how a non integrated plot infers more high end, but by the same logic it would imply more low end, because it's imposing -6dB per octave roll off at all frequencies. 

Fishman provided a bode plot for marketing their product, which looks like it has been touched up a little, and it looks like their chart indicates that they have a flat response to the resonance, then a -6dB slope past the resonance.  An integrated plot of a real guitar pickup will show a pickup having a -12dB/oct slope pas the resonant peak, and the integrated measurement of a Fishman Fluence conducted here also shows that expected measurement closer to -12dB/oct past the resonance. It seems to me that the Fishman Fluence is responding like a real passive pickup, and it it weren't, that would cause it to sound less like a passive pickup, and more like an active one. I have a Fishman Fluence bridge pickup in a Schecter, and I can confirm that it's response reminds me of an uncovered PAF, almost to a fault, I wish it provided more treble in one of it's two voice modes. 

[aquin43]

Removing the integrator adds a uniform 20dB/decade rise so there will be less low end and more high.

Since the plots are log-log, it is easy to add lines at standard slopes using a drawing program such as inkscape. This allows easy estimation of roll-off rates e.g. (labelled dB per dec / dB per oct)

[antigua]

Jul 24, 2022 3:58:50 GMT -5 aquin43 said:

Removing the integrator adds a uniform 20dB/decade rise so there will be less low end and more high.

Since the plots are log-log, it is easy to add lines at standard slopes using a drawing program such as inkscape. This allows easy estimation of roll-off rates e.g. (labelled dB per dec / dB per oct)

I understand this, but I still suspect the Flishman tech is incorrect about the treble response, since his comment regarded treble but said nothing of the bass response. I think the integrated plot is the context in which the Fishman tech is speaking about.

[gpdb]

To provide some context around my contact's response of "there's more high end," I don't believe it was necessarily a scientific statement but a comparative one. He first asked to add another PAF bode plot next to it, and then made the comment that it should have a shallower roll-off than standard passive pickups. My assumption is that since Fluence pickups determine their own curves through programming, I would assume his statement reflects that they intentionally add more high end than standard pickups to create better clarity. Whether it's placebo or not, you can find tons of comments online about people speaking to the "clarity" of Fluence pickups. I've never A/B tested them back to back, but they sound great. He also stated that the clarity had nothing to do with the signal being buffered. Even if the high end is off, the larger plot that seems off is the Neck voice 2. If that curve was correct, I would assume it would be entirely unusable. I never personally liked voice 2 of the neck, but I don't recall it being that crazy.

[antigua]

If your bode plotter says the Fluence and the regular HB both have a -12dBV drop for each doubling of frequency, then it is what it is. The pickup will not have a different transfer function when it's in the guitar, versus when it's under the test coil. 

I have a guitar with the Fluence, and the vintage / modern push pull, and the vintage voice has the characteristic -12dB roll off, which showed up in my plot also:
 
If people are  saying it sound clearer, I think it might be because it's very noise-free, very quiet. I'd go so far as to say I think it's more noise-free than EMGs. 

Note that the modern voicing has a softer roll off, you do hear more treble response with it, but it's somewhat overpowered by the mid-hump below 1kHz. I kind of wish the modern voice was flat, but they seemed to have been responding to the heavy metal market and had copied the 80's EMG response profile.